ABSTRACT
OBJECTIVES: A prior single-center study demonstrated historical and exam features predicting intracranial injury (ICI) in geriatric patients with low-risk falls. We sought to prospectively validate these findings in a multicenter population. METHODS: This is a prospective observational study of patients ≥65â¯years presenting after a fall to three EDs. Patients were eligible if they were at baseline mental status and were not triaged to the trauma bay. Fall mechanism, head strike history, headache, loss of consciousness (LOC), anticoagulants/antiplatelet use, dementia, and signs of head trauma were recorded. Radiographic imaging was obtained at the discretion of treating physicians. Patients were called at 30â¯days to determine outcome in non-imaged patients. RESULTS: 723 patients (median age 83, interquartile range 74-88) were enrolled. Although all patients were at baseline mental status, 76 had GCS <15, and 154 had dementia. 406 patients were on anticoagulation/antiplatelet agents. Fifty-two (7.31%) patients had traumatic ICI. Two study variables were helpful in predicting ICI: LOC (odds ratio (OR) 2.02) and signs of head trauma (OR 2.6). The sensitivity of these items was 86.5% (CI 73.6-94) with a specificity of 38.8% (CI 35.1-42.7). The positive predictive value in this population was 10% (CI 7.5-13.3) with a negative predictive value of 97.3% (CI 94.4-98.8). Had these items been applied as a decision rule, 273 patients would not have undergone CT scanning, but 7 injuries would have been missed. CONCLUSION: In low-risk geriatric fall patients, the best predictors of ICI were physical findings of head trauma and history of LOC.
